Visual walkthrough — Modal analysis — natural frequencies, mode shapes

This page rebuilds the central result of the parent topic — the eigenvalue problem that hands us a structure's natural frequencies and mode shapes — starting from two masses on springs and nothing else. Every symbol is earned before it is used. Follow the red and amber arrows in each figure; the pictures carry the argument.

Step 1 — Two masses, two springs: the whole universe

WHAT. We build the simplest structure that can still show interesting vibration: a wall, a spring, a mass, another spring, another mass.

WHY. One mass on one spring bounces in a single boring way. Add a second mass and the two can either move together or fight each other — that "together vs. against" is the seed of every mode shape in a rocket. We start here because it is the smallest system with more than one answer.

PICTURE. Below, $m_1$ and $m_2$ are the two masses. The symbols $x_1$ and $x_2$ are just arrows measuring how far each mass has slid to the right from its resting spot. Positive means right, negative means left. That is all $x$ ever means.

This is the discretised skeleton of any structure — see Multi-degree-of-freedom systems for how a real spacecraft becomes thousands of such blocks via Finite element analysis (FEA).

Step 2 — Newton's law on each mass → the matrix equation

WHAT. We write $F=ma$ for each block, counting every spring pulling on it, then stack the two lines into one tidy box of numbers.

WHY. A spring's force is $-k\times (\text{its stretch})$ — the minus sign says it always pulls back toward rest. Mass 1 feels two springs (the wall spring stretched by $x_1$, and the middle spring stretched by $x_1-x_2$). Mass 2 feels only the middle spring. Writing both and stacking them lets us treat the pair as one object.

PICTURE. The figure shows each spring's stretch as a coloured segment and the force arrow it produces on each mass.

Mass 1: the wall spring is stretched by $x_1$; the middle spring is stretched by $(x_1-x_2)$.

Mass 2: only the middle spring acts.

Now stack them. We invent two grids of numbers so the two equations fit on one line:

Step 3 — Guess the motion: everything pulses at one rhythm

WHAT. We guess that in a natural mode, every mass rides the same sine wave in time, differing only in how big its swing is.

WHY. With no damping, a spring–mass system rings sinusoidally — that is the one motion whose acceleration points straight back toward zero, which is exactly what springs demand. So we try a solution where the shape (who moves how much) is frozen and only the amplitude breathes in and out.

PICTURE. Two masses, one shared clock. The tall amber curve and short cyan curve share the same period; only their heights differ. The frozen height-ratio is the mode shape.

• $\{\varphi \}$ (say "phi") — the mode shape: a list of relative amplitudes, e.g. $\left[\begin{array}{c}1\\ 1.6\end{array}\right]$ means "when mass 1 moves 1, mass 2 moves 1.6." It carries no units and no time — pure pattern.
• $\omega$ (say "omega") — the angular frequency in rad/s: how fast the shared clock ticks. This is the natural frequency we are hunting.
• $\theta$ — a phase offset (where the sine starts); it will cancel and never matters.

Because $\sin$ curves back on themselves, differentiating twice in time gives the same shape times $-{\omega }^2$:

The $-{\omega }^2$ is the fingerprint of sinusoidal motion: fast pulsing (big $\omega$) means fierce acceleration.

Step 4 — Substitute and cancel the clock → the eigenvalue problem

WHAT. Plug the guess into Step 2's equation and cancel the sine that every term shares.

WHY. The sine wave $\sin (\omega t+\theta )$ multiplies every term identically, and it is nonzero at almost every instant, so we may divide it out. What survives is a pure statement about shapes and frequencies — time has been squeezed out entirely.

PICTURE. The clock $\sin (\omega t+\theta )$ lifts off both sides like a common factor peeling away, leaving a timeless equation about the shape $\{\varphi \}$.

Substituting $\{\ddot{x}\}=-{\omega }^2\{\varphi \}\sin (\cdot )$ and $\{x\}=\{\varphi \}\sin (\cdot )$:

Divide out the shared $\sin (\cdot )$:

Why must a determinant vanish? We do not want the boring answer $\{\varphi \}=\{0\}$ (nothing moves). A nonzero shape can only survive multiplication by the matrix $([K]-{\omega }^2[M])$ into zero if that matrix is squashing — mathematically, if its determinant is zero:

Step 5 — Turn the determinant into a polynomial and solve

WHAT. Compute that determinant for our 2×2 numbers, get a polynomial in ${\omega }^2$, and solve it.

WHY. The determinant condition is the single equation that filters which $\omega$ are allowed. Expanded, it is an ordinary quadratic in the variable ${\omega }^2$ — and quadratics we can always crack with the quadratic formula.

PICTURE. The parabola in ${\omega }^2$; its two crossings of zero are the two allowed ${\omega }^2$.

Expanding term by term:

Treat ${\omega }^2$ as one unknown; the quadratic formula gives

Two masses → two natural frequencies. (In general $n$ masses give $n$ frequencies.)

Step 6 — Recover the two shapes

WHAT. Put each ${\omega }_i^2$ back into $([K]-{\omega }_i^2[M])\{\varphi \}=\{0\}$ and read off the amplitude ratio.

WHY. Each allowed frequency squashes the matrix in exactly one direction — that surviving direction is the mode shape. We fix ${\varphi }_1=1$ and let the equation tell us ${\varphi }_2$, because only the ratio is physical (Step 3 froze shape, not size).

PICTURE. Mode 1 both masses swing the same way (amber, in-phase); Mode 2 they swing opposite (cyan, out-of-phase) with a still node between them.

Low frequency, ${\omega }_1$: first row $2k{\varphi }_{11}-k{\varphi }_{21}=\frac{(3-\sqrt{5})k}{2}{\varphi }_{11}$ gives

High frequency, ${\omega }_2$:

Step 7 — The shapes are orthogonal (the payoff)

WHAT. The two mode shapes obey $\{{\varphi }_1{\}}^T[M]\{{\varphi }_2\}=0$ and $\{{\varphi }_1{\}}^T[K]\{{\varphi }_2\}=0$ — a matrix-weighted right-angle.

WHY. This is what makes modes useful: they are independent building blocks. Any wobble of the structure is a sum of these two, and — because they are orthogonal — each evolves as its own separate single-mass oscillator. That is how a monstrous coupled system melts into simple pieces we can solve one at a time. Full proof and consequences: Orthogonality of mode shapes.

PICTURE. In the $M$-weighted picture the two shape vectors sit at a clean right angle; projecting the structure's motion onto each axis reads off "how much of that mode is present."

Using $[\Phi ]=[\{{\varphi }_1\}\{{\varphi }_2\}]$ and $\{x\}=[\Phi ]\{q\}$, pre-multiplying the equation of motion by $[\Phi {]}^T$ turns the coupled matrices diagonal:

• $q_i$ — the modal coordinate: how strongly mode $i$ is switched on right now.
• $M_i^{\ast },K_i^{\ast }$ — the mode's own effective mass and stiffness, with ${\omega }_i=\sqrt{K_i^{\ast }/M_i^{\ast }}$.
• $Q_i=\{{\varphi }_i{\}}^T\{F\}$ — how hard an external force (like a launch shake) pushes that mode.

The danger, made precise: if a launch load contains energy at ${\omega }_i$, that one decoupled oscillator resonates. Modal analysis pins down every ${\omega }_i$ so designers can dodge it or damp it — see Structural damping mechanisms and Frequency response functions.

The one-picture summary

From a wall, two springs and two masses, we wrote Newton's law (arrows), guessed a shared sinusoidal rhythm, cancelled the clock to reach $[K]\{\varphi \}={\omega }^2[M]\{\varphi \}$, solved a quadratic for two frequencies, read off two orthogonal shapes, and decoupled the whole thing into independent oscillators.